Abstract 1212P
Background
Severe irAEs caused by ICIs affect treatment efficacy and benefit. Current research on irAEs is mainly focused on early prediction, with a lack of near-term prediction. Studies have reported that baseline neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR) and absolute eosinophil count (AEC) can predict irAEs. Our aim is to explore the near-term predictive value of NLR, PLR, and AEC for grade 3 or higher irAEs caused by PD-1 inhibitors.
Methods
Data were collected from cancer patients treated with PD-1 inhibitors in our department from January 2020 to May 2022. NLR, PLR, and AEC data were collected one cycle before and during the occurrence of irAEs (median cycle number was 2nd and 3rd cycles, respectively). Logistic analysis was used to analyze the correlation between NLR, PLR, AEC, and irAEs, and to construct a prediction model. The model's performance was assessed by obtaining sensitivity and specificity through the ROC.
Results
Out of the 138 cancer patients, 47 experienced grade 1-2 irAEs, and 18 had grade 3 or higher irAEs (including 2 fatal irAEs). Multivariate analysis showed that the current cycle's NLR (OR, 1.839, p=0.000) and PLR (OR, 0.994, p=0.029) were independent risk factors for irAEs. Model A predicted the occurrence of irAEs in the next cycle with an area under the curve (AUC) of 0.788, a sensitivity of 69.2%, and a specificity of 68.5%. When the Model A was ≥38.8% (cutoff value), Model B (AUC =0.900) was entered, predicting the occurrence of grade 3 or higher irAEs in the next cycle with a sensitivity of 67.7% and a specificity of 72.6%. Similarly, Model C predicted the occurrence of irAEs in the current cycle (AUC=0.865) yielded a sensitivity of 81.5% and a specificity of 86.3%. When the Model C was ≥52.8%, Model D was entered, predicting the occurrence of grade 3 or higher irAEs in the current cycle (AUC=0.888) with a sensitivity of 83.3% and a specificity of 83.0%.
Conclusions
The model consisting of NLR, PLR, and AEC in this study can predict the occurrence of grade 3 or higher irAEs within one cycle. Compared with early prediction, our near-term prediction model maximizes the number of cycles of immune therapy obtained by patients, thereby improving the effectiveness of immunotherapy, which has significant clinical value.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.
Resources from the same session
1251P - Development of a deep learning algorithm for lung cancer diagnosis using methylation and fragment size profiles of cfDNA
Presenter: Jiyoung Huh
Session: Poster session 14
1252P - Quantitative cell signaling activity profiling of solid tumors to support personalized treatment in the FINPROVE basket trial: Presentation of skin tumor data
Presenter: Diederick Keizer
Session: Poster session 14
1253P - Analytic validation and implementation of OncoDEEP: A pan-cancer comprehensive genomic profiling NGS assay for assessing homologous recombination deficiency (HRD)
Presenter: Marcel Trautmann
Session: Poster session 14
1254P - Retrospective analysis of brain OMX: Diagnostic tool for structural (T1) and functional connectome (RS-FMRI) analysis of brain
Presenter: Swarnambiga Ayyachamy
Session: Poster session 14
1255P - Evaluating GPT-4 as an academic support tool for clinicians: A comparative analysis of case records from the literature
Presenter: Marcos Aurelio Fonseca Magalhaes Filho
Session: Poster session 14
1256P - Value of detection of peripheral blood circRNA based on digital PCR in the diagnosis of lung adenocarcinoma
Presenter: Jihong Zhou
Session: Poster session 14
1257P - Double heterozygous prevalence in hereditary cancer syndromes in Northern Mexico population
Presenter: Carlos Burciaga Flores
Session: Poster session 14
1258P - Does FDG PET-based radiomics have an added value for prediction of overall survival in non-small cell lung cancer?
Presenter: Andrea Ciarmiello
Session: Poster session 14
1260TiP - Enhancing lung nodule discrimination with a novel cfDNA test: The cancer signature ensemble (CSE) approach
Presenter: Young-Chul Kim
Session: Poster session 14
1773P - ICECaP-2: Validation of metastasis-free survival (MFS) as a surrogate for overall survival (OS) in localized prostate cancer (LPC) in a more contemporary era
Presenter: Wanling Xie
Session: Poster session 14